If you like what you heard on this week’s episode, be sure to give us a Review, Like, or Share to help support us!
Did you know that mastering Chatbots doesn’t have to be complex?
To show you, we’ve interviewed three Chatbot experts to give you their opinion and viewpoint on how to be successful with Chatbots.
From scaling to fine tuning, we hope you enjoy this deep dive.
Subscribe to BoostSauce to learn even more marketing and sales knowledge, faster…
When it comes to Chatbots, there are only a few people we turn to for amazing advice that works across the board for different goals.
Whether you’re in SaaS, eCommerce, or lead gen, you’ll be excited to learn that the recipes these experts will share will all help you hit your goals faster.
In Order Of The Guests Below:
Mark Kilens is VP of Content and Community at Drift where he leads the blogging, editorial, social, Drift Insider, and HYPERGROWTH teams.
Prior to joining Drift, he served as VP and founder of HubSpot Academy. As an enthusiastic former customer, Mark joined HubSpot in 2010 as an Inbound Marketing Consultant, where he spent two years working with thousands of HubSpot customers to scale their inbound strategy.
From there, he went on to build HubSpot Academy from the ground up, and in doing so, educated millions of people and created a best-in-class hub for marketing and sales content and training.
Nuggets Dropped x37
“Start by testing bots on high traffic (high-intent) pages first.”
Larry Kim is the CEO of MobileMonkey, a provider of Facebook Tools for messaging. He’s also the Founder of WordStream, the World’s Largest PPC Marketing software company, employing over 300 people and managing a billion dollars of ad spend for tens of thousands of customers worldwide – acquired by Gannett in 2019 for $150 million. Larry is a contributor to Inc Magazine and CNBC. In 2019, he was recognized by the United States Senate and the US House of Representatives for outstanding entrepreneurial contributions to the community by recent immigrants to the USA.
Nuggets Dropped x62
“Learn who the user is/what they want in smaller, private convos.”
In early 2015, Johnathan Dane left his first PPC company in Utah and moved back to his hometown in Southern California. What was the first thing he did when he got back? He started another PPC company and he called it KlientBoost. He started with no clients, no office, and a one-person team. Armed with some savings from the first company and a designer he convinced to join him, his dream was to build a PPC SaaS company (his first PPC company was an agency).
Nuggets Dropped x11
“Always remember that there’s another human across the screen.”
Chatbot Mastery With Mark Kilens
Johnathan: Hey, Mark, Jonathan over here from BoostSauce. How are you?
Mark: I am doing well. How are you?
Johnathan: I’m pumped. I’m actually super excited to chat with you. We’ve been following Drift for quite a bit outside and your VP of Marketing is a really cool dude, too. So I’m pumped to have you on the show. So you’re basically the VP of Content and Community at Drift. What does that role entail?
Mark: Good question. So we have taken a slightly different approach with content and community at Drift. Instead of it being kind of siloed across the organization, we have like one holistic content strategy and the same goes for community.
So I’m on the marketing team. I work hand-in-hand with Dave Gerhardt, Kate Adams, who runs our demand gen, and I’m focused on building content for marketing, but also, for sales, the customer success team, and running community programming including the Hypergrowth event for all of Drift too.
Johnathan: Dang, so like across all divisions, it seems like, not just for marketing which I feel like is kind of the norm.
Mark: Yeah, no, across everything because you wanna really make a really personal end-to-end customer experience for everyone who interacts with the Drift brand and the business and, as you know, content is so important to all that.
Johnathan: Right, that’s crazy. Okay, cool. Well, we have you on the show. We wanna talk about chatbots. I mean, that’s a lot of what Drift is today and I know you guys are doing more and coming out with more stuff too.
But I wanna be very tactical and talk about how marketers or, you know, people, if they don’t wanna feel like they’re calling themselves marketers, can use Drift to their advantage. So if I were to ask you like, you know, give me the rundown of Drift and why it’s useful right now.
Mark: Sure, Drift can help in many different ways. The main way I think about what Drift can deliver for a business is in outcomes. The outcomes that we help deliver are more qualified leads, really help you book more meetings to generate more pipeline, and grow your revenue.
So let’s break that down a little better, alright? So the first thing we do is we help you understand how to start more conversations with people visiting your website.
Mark: That is the most important piece. Then, we help you actually have those conversations. So we have this framework that we should probably start with called The Conversational Framework. It’s three parts. It’s engage, understand, and recommend.
And we use that framework to both design how you wanna start a conversation with someone using a chat bot. And then, actually, we use that framework to help a sales development rep, a business development rep, actually have a productive conversation that can then result in a meeting for an account executive.
Johnathan: So it’s like Drift is asking questions in advance so the actual SDR or BDR has that information before they jump on that scheduled meeting, for example?
Mark: Yeah, this is two ways to do it. One is like, you know, bot to human. So like chatbot to human, where the chatbot is gonna help do a lot of that understanding, you know, also known as what you were saying, qualifying the buyer, or disqualifying, or routing them.
Or, it’s just human to human where, you know, you have people, we sometimes call them conversational development reps. But someone who’s actually online, ready to engage with people that are coming to the website either through an inbound or an outbound type of motion. So that’s more the human to human.
But we definitely recommend, unless you have a really specific set of target accounts you’re going after, which, you know, more and more companies do these days using an account-based marketing approach. You definitely need bots across your website. And you should use these bots on your most trafficked, high-intent pages to start with. That’s the key.
And then, we, that’s the key there. Yeah, I mean, the blueprint, yeah, the blueprint we developed will help with this too. We have a blueprint called The Conversational Marketing Blueprint which is free, you can use it for anything.
But the way we would like you folks to try using it for is how do you start conversations with people in the right way. So it’s broken down into five different parts?
Do you want me to go through the five parts?
Johnathan: Yes, please.
Mark: So it starts with what page you want to start a conversation with someone on. So that’s the first step: What page? And again, I recommend starting with high-intent, highly-trafficked pages. Your home page, your pricing page, your demo page, your Contact Us page, your product page, et cetera.
Then, it’s like, who do you wanna have a conversation with? The bot could show up for anyone visiting that page, or you could get really, really specific and target only your customers that are sales, or your buyers, I should say, that are sales-engaged.
Johnathan: Okay, how do you define that?
Mark: Yeah, I mean, anyone who fits into your ideal customer profile, who has shown intense, both, on the behavior side and maybe demographic side. So using different types of intent signals. I think you definitely should come up with some type of way to score people coming to your website.
Mark: Lead-scoring is one way to call it. But you’ve got to use intent signals.
Mark: So that’s the Who, right? Figure out your Who. And that’s gonna be done through a combination of, maybe, you integrating Drift to your CRM. You’re working on an automation platform. But you being really specific.
Then, it’s like, does it matter where they came from? So they’re on a page. So what page are they on? Who are you trying to start a conversation with, and does it matter where they came from? It could be from anyplace. It doesn’t matter where they came from.
Or, you wanna show a specific bot, because they came from a specific nurture campaign. So you wanna engage them. You wanna design a conversation that is really specific to that destination they came from and who you’re trying to talk to.
Johnathan: Got it, got it.
Mark: You’re really specific about it, right? I mean, it’s really advanced but you wanna get there, right? But you can keep it simple to start with.
And then, it’s like Why? This is key, Jonathan. Why? Why does the person who’s visiting that website page wanna have a conversation with you? And then, why do you wanna have a conversation with them?
Mark: Yeah, that’s important to think about. So why would someone go to your pricing page, do you think?
Johnathan: Because they are intrigued to potentially work with us, but they’re concerned that the pricing might be too high. And so, that’s why they’re there.
Mark: Yeah, definitely. They definitely are probably interested in using your product in some way, or your solution in some way. So why would you wanna have a conversation with them, though?
Johnathan: Because there’s that intent now that they’re on the pricing page which is beyond other pages that are on our site, too. So they have some engagement already and, if they have intent and engagement, you know, our goal is to get them as a client if they’re a good fit.
But we can’t know that on the website just yet. So that’s why we’d wanna have a conversation with them.
Mark: Yeah, you wanna help them, you know, help them with some questions, maybe. See if they are really kind of further down the buying journey. So, 100%.
So the Why’s important and the last W is simple. This is actually When. It’s When the buyer or that website visitor decides to engage with the bot.
Mark: So that really, if you use those five W’s, you can build any type of bot you want. It could be a really simple bot for a homepage. That’s like a conversation starter bot that’s basically targeting anyone, really low targeting. It can really be anyone.
Or you could get really specific with those five W’s and engage with a specific list on a specific set of pages from a specific campaign to increase the chance that they’re gonna start a conversation with you.
Johnathan: And you guys, I don’t know if you still call it this, but are they called playbooks, the different types of logic the bots can take? Is that correct?
Mark: Yeah, it’s playbooks. That’s the primary feature we have in the Drift chat tool. And what we have, actually, the playbook marketplace that comes prebuilt with about 10 different playbooks that are gonna help you actually design the best conversation.
Because, there’s two components to being really good with Drift and conversational marketing. There’s building the bot, which is what we just went through, and then, it’s designing the conversation which is that Engage, Understand, Recommend framework I was mentioning.
Johnathan: Right, okay, so there’s two parts to it. I like that.
So to your first part of like the five W’s, I’m not gonna mention them back because I already forgot them, but what we’ll basically do is, you basically can have a lot of different scenarios depending on the type of page they’re on, where they came from, like you said, too, if there’s additional data from a lead-scoring perspective, you can have that figured out.
And then, we’re getting into the actual conversation. So take us through that. What do you do there?
Mark: Yeah, a lot of people fail to design a conversation that matches with the five W’s. They either do one or the other and they don’t do both together. And, if you do both together, you can have, you know, easily 5% or 10% engagement rates with your bot.
Mark: We see, on average, like a 3% to 4% to 5% engagement rate in terms of our customers. That’s what they see. But on your website, you can receive up to 30%-35% engagement rates.
Mark: So yeah, I mean, it’s crazy, right? So what you wanna do is you wanna first ask yourself, what question is gonna be the most engaging question, which, again, is the first part of the framework, that is gonna pull someone into a conversation?
Mark: And that’s dependent on the Who. Who are you trying to start that conversation with? And it’s dependent on what page and where they came from.
Johnathan: So this is, I’ve got a quick question for you because I’ve played, so this is back in the day when I was playing with Olark. You know Olark, the chat software?
Johnathan: Before you guys came alive, that’s what we had to use as marketers. So I remember Olark back in the day, that we had what they call chat greeters. And there was no logic, there was no playbook. There was definitely not any conversational intelligence, whatsoever.
But I remember, and you might be alluding to this. So if you wanna give me a chicken sound back, maybe you can. But, basically, the questions that we found were the best to start any kind of engagement with the chat tool were really simple questions to ask. Meaning like something that was an easy numerical answer back, or something that wasn’t super-threatening.
Because what I saw a lot was people using, “Hey, welcome to our site. “Thank for hanging out. “Please ask me a question if you have one.” And I related that to a person that works in a clothing store. You always say, “No thanks, I’m just looking. “Get out of my face.”
But the small, easy-to-answer questions were the ones that got the quickest amount of engagement, and then, also, lead volume. So, sorry, I don’t wanna steal your thunder. Continue.
Mark: No, cuh-caw. That’s my chicken sound. Cuh-caw.
Johnathan: Thank you.
Mark: That’s good. No, definitely. So let me give you some examples. So if you’ve got a bot on the homepage and it’s targeting first-time visitors to the website, and you know that because they don’t have cookies yet.
One of the best questions you could ask to qualify them, this is more on the qualification side, is take the value proposition that’s on your website, or the thing that you help people do, the solution you help people implement, and ask that in the form of a question.
Johnathan: Got it.
Mark: For Drift, it’s like, “Hey, just curious, would you like to book “more meetings for your sales team?” Or like, “Hey, would you like “more qualified leads every month?”
Mark: And then, the pro tip here is, “Yes, I would”, right? That’s one of the fun responses. Or, “Maybe, I’m not sure.” It’s always a double-yes question. It’s never a yes-no.
Johnathan: Ah. I like that. Okay.
Mark: A key. And then, that’s the engage key.
Mark: That’s the key to understand, is when the bot’s gonna ask maybe one, maybe two, maybe three or four follow-up questions to better understand that website visitor, to then lead to the best possible recommendation for them.
Johnathan: Got it, okay.
Mark: So those questions are definitely gonna be more about qualifying a little bit. You know, it should be set up so that you don’t ask a visitor the same question more than once, which we’ll help you do.
But, really, the questions the bot’s gonna ask, they should be done in a conversational way. So one of the things that everyone should do when they build bots is literally read out the bot flow with a friend or a coworker and have that back and forth and see if it sounds really human and really personal.
Johnathan: Okay, so is this abbreviations and stuff, or what do you recommend with that?
Mark: Oh, yeah, it could be abbreviations. It could be how you talk. It could be not in complete sentences. It could be with emojis. You could use a gif in the bot.
It gets a little bit more advanced, but you wanna bring in the same type of messaging experience that sticks within your brand into your bots. Because, again, that’s how many, so many people these days use messaging to communicate. We all do, right?
Mark: So you wanna make it feel like you’re having that kind of conversation with one of your friends of family members with this company’s bot, again, aligned to the business’s brand.
Johnathan: Right, for sure. That makes sense. I like that. Okay.
Mark: So, yeah, that’s the understand piece, and the recommend piece is really, two things are gonna happen. You know, one is you’re gonna recommend an idea. You might recommend a piece of content. You don’t always have to link them to something.
In Drift bots and playbooks, you could actually have a link that unfurls so they see the video, or they see that content right in the bot flow. With Drift, that’s like, the bot’s gonna share something with them or the bot’s gonna schedule time for them to talk to someone in the future.
Or, it might recommend, “Hey, let’s have you speak to someone right now,” because the person’s like, “Yeah, actually, I had questions about pricing,” and, “Yes, I am free to talk. I’d like to speak to someone about pricing right now.”
Johnathan: Right. It makes perfect sense.
Mark: And the great thing about bots, in my opinion, versus forms, is bots have optionality. A form is linear. It’s one thing. You’re like, this is the thing you’re gonna get and you’re probably gonna have to wait to get it.
And you probably have answered these same questions before. And it’s just not personal, there’s no options, it’s boring. A bot is the complete opposite of that, right? There’s options, you can have personality, it’s interactive, it’s just so much better.
Johnathan: Right, completely. And I’m thinking of our own form, which is a three-step form on our website where we start with a small ask and it progressively gets more intense. But there’s nothing holding us back from putting the whole thing on Drift, basically, in regards to all the different angles that we’re not asking where people also feel like they’re getting a more custom-tailored solution.
Whereas, like you said, with a form being linear, they know, at the end of this, that they’re probably gonna have to talk to a salesperson. And they still might, but they might have a higher incentive now to wanna do that, because they’re giving the answers to Drift and, in turn, that made more sense for them.
Mark: 100%. I just think a form is very limiting these days. And sometimes, a form might still be useful, but it’s not often that it’s useful. You really wanna, yeah, you wanna use a bot so that you can actually ask them, them being the website visitor, better questions to help them with whatever they need, right, and to tailor that experience, and a form can’t do that.
Johnathan: Do you guys use any more on the Drift site right now?
Johnathan: No, alright. I was gonna see. Good. You’re living and breathing while you’re preaching. Good, cool.
Johnathan: So, obviously, a lot of things are making very, very crystal-clear as far as the framework and the fundamentals of setting things up.
Do you have any favorite greeter questions that you’ve seen work extremely well from an engagement perspective, or is it all specific to the, you know, the five W’s that we talked about already?
Mark: There’s definitely some. I mean, it’s really specific to the page. That’s where it all starts.
Johnathan: That’s the best one.
Mark: Yeah, the page is super-important, and also, really, the who. The where doesn’t matter as much. I mean, that’s where it can get really sophisticated. You know, the page and the who.
So, for example, I recommend you at least know the difference between your customers, right? So you have a market foundation platform. We sync with all the big ones, and we sync with the big CRMs, we integrate with them.
You know, you don’t wanna be asking your customers questions to understand if they’re a buyer. That’s a fact. So you wanna show, you either don’t wanna show the bot, right? That’s always an option. Just don’t show the bot to those people.
Or have it set up so that the bot is gonna engage with that customer and do it in a way that’s gonna lead them to maybe getting some support help, or asking them how they’re feeling or experiencing your product or service.
So, for me, the questions for the buyers all come down to what page and who? So let me give you an example. One is the value prop. That’s a slam dunk.
Another one that’s actually really interesting is just asking them who they are. So this goes to the persona question. You can ask them, basically, “Oh, hey, are you a property manager looking to increase your rental efficiency?”, right? Something like that, for example. Kind of the value prop and the person, who they are–
Johnathan: The title of them. Not like, you are Mark Kilens. It’s specific to their title.
Mark: Exactly, exactly. You know, you go up to them. “Hey, how can I assist you with our pricing and packaging options? Because that’s a lot to choose from.” That’s, again, very specific to the pricing page.
But something like that where they go on the pricing page, I guarantee, there’s a couple, if not more than a couple of people out of a hundred who are on that page that have questions and would like assistance.
Johnathan: For sure.
Mark: You know what I mean?
Johnathan: For sure, absolutely.
Mark: So I also think you don’t have to overthink it. Just don’t overthink it.
Johnathan: Okay, I like that. So, so far, we talked about all the different types of bots and playbooks you can use depending on the location that they’re on on your site, where they came from. And I’m gonna get a summary for everybody else because there’s a lot to understand here, which is awesome. That’s Step One.
And then, the actual side of the engagement portion. Do you feel like there’s a lot more value in split-testing the initial question versus further down in the logic of the chatbot? What has the biggest impact from a result perspective?
Mark: It’s really depending on the outcome you’re trying to drive. And, most of the time, it’s booking a meeting or qualifying that lead. And, in most cases, it’s actually both. You’re doing both at the same time.
Mark: So I think you’ve got to get a page. I know you’ve got to get a couple of bots on your website, on these high-intent pages that seem to get a lot of traffic. And then, give it at least, you want, minimum, a thousand people to see it.
Mark: And then, you can start playing around a little bit. But I honestly think, if you can give it, you know, say, a couple of weeks, a couple thousand people have seen it, that means you probably have a 3%-5% engagement rate.
And it’s pretty easily done with the playbook that we have in the marketplace. You can then start tweaking it a bit. But I almost would say, instead of tweaking those bots that are working decently well, expand it to other pages.
Johnathan: Got it. That makes sense. I like that. Just because you’ll get more volume as you go along. It’s not just about increasing the percentage, which you can still do, but I also think you learn a lot faster by seeing if it works on other pages on your site, as well.
Johnathan: Got it. Cool. Well, man, this has been super, super-helpful. I’m loving the starting fundamentals, and also, just understanding the structure behind it, the skeleton and the anatomy of being successful with Drift.
If we have any questions, man, we’re gonna definitely follow up with you. But, Mark, I wanna thank you so much for your time, man. Really appreciate it.
Mark: It’s my pleasure. Thanks for having me.
Johnathan: Awesome, thank you so much. Talk soon.
Mark: Alright, see you.
Chatbot Mastery With Larry Kim
Johnathan: What’s up, Larry Kim? How are you?
Larry: Hey Jonathan, I’m doing great, thanks for having me today.
Johnathan: Yeah, of course, man. So you and I go a little bit back. We’ve been on stage together, we’ve done a webinar together, now we’re doing a podcast together. Just a lot of memories between you and I. So I appreciate you being here.
Larry: Awesome. So many different channels, yeah.
Johnathan: True, true. So, you’re one of the best people I know when it comes to the Facebook Messenger side and the chatbot abilities of what you can do there. You’ve obviously left your previous company and now founded MobileMonkey, and I was really impressed with what we went through and looked at before, and we’ll even link to that webinar in our show notes.
But I wanted to take the listeners through how you get started with anything in that world even. Like, the foundational stuff, all the way through to how they are able to scale it. So if you’re talking to somebody who, let’s say, has a decent brand presence, they have a community, a lot of things are going well for them, but they wanna get a little bit better and a little bit more tactical.
Where would you guide them when it comes to what MobileMonkey does, but also what the Facebook Messenger side does, as well?
Larry: Right, well, personally, I think a place to start is, a lot of people are looking at chatbots as a kind of a next generation lead ad. So, I’m sure everyone’s familiar with lead ads and how the other things work.
Instead of clicking a button that goes to a web page like clicking on an ad on Facebook, it pulls up a form, a form with like you know, your email address and your phone number.
Johnathan: Right, right.
Larry: and stuff like that. Now, the problem with that ad format is that it just ends up being a kind of dumping ground. It’s like an email storage tank and these emails, they just kind of decay.
Larry: And so one of the things that a chatbot can do on Facebook, and you know there’s tons of different use cases, but this is a common one for marketers, is to just, instead of using a kind of lead ad to give a Messenger ad.
And so the way that works is just when they click on the Facebook ad, instead of sending them to a web page, it sends them into a conversation with a chat, like with your business.
Larry: Like through messaging. And you can respond to that manually, but that’s no better than email stuff for you.
Johnathan: Right, right.
Larry: Just dumping these emails into a storage tank and you’re getting back to them later. Where it gets very interesting, is that you can connect all these chatbot automations, to the kind of a chat landing page, where your sending people to when they click on your ad.
And so, it’s quickly what you do is you just design these conversational funnels. You try to understand who the user is and what they care about by asking them a series of questions and then really, you know, then trying to direct them to the correct offer.
Larry: You know and on Facebook ads, you can have an audience of 100,000 people, or a million people, and you’re kind of making assumptions that, “I think that these million people are gonna really be interested in this particular offer.”
But in reality, you’re just assuming that that’s the case. You’re kind of limited because you only have the ability to make one offer.
Larry: You know? But in messaging we can kind of find out, what exactly that they’re interested in.
Johnathan: Sure. More like a menu, so to speak.
Larry: Yeah, they can look through it like a little questionnaire. And send them to the correct answer. And because it’s new and it’s mobile friendly and you know, people really love this.
People kind of view chat as a very intimate space, where they converse with their family and their friends and so, set them cleverly, like, kind of I would say, you know, go a little off brand, be a little bit more, you know, informal. In terms of like the images that you send them.
Larry: And the, language, like I never use capital letters. Like emojis.
Larry: And I send GIFs and all this stuff. If you can go off brand a little bit, what you’ll find is that actually building kind of out programmatic brand affinity. Like getting like they’re kind of building a relationship with this brand by chatting with them, even though it’s just a robot answering the questions and asking questions.
Johnathan: It’s funny you say that, because it reminds me of our first early days of KlientBoost, where we would cold email people prospects and I would have a GIF in there. You called it a GIF. That is gonna be a whole separate episode of like where we’re gonna figure that out.
But basically, it was a guy with a cowboy hat, rodeo riding a pepperoni pizza slice and like swinging his arm in the air, as if he has like a lasso, in space, basically. And that alone, was obviously me not taking myself too seriously and wanting to be different.
The thing that you mentioned, which I thought was really cool is, when people do a lead ad, or when people convert on a form, on a landing page, they know they’re not gonna get an answer to their question right then and there.
But because of the functionality of a chatbot, they feel as if like the journey that they can go through and the path of logic, depending on how you set it up, even though it’s automated, can get them closer and closer to have some of those questions answered immediately and so, it’s a way better experience for them.
It’s way easier for you, because you’re also gonna be able to have more qualified people, most likely too, that are already further down the journey, with questions answered already. Compared to your other type of advertising format. Did I get that right?
Larry: Yeah, absolutely.
Johnathan: Okay. I’m gonna give myself a nugget.
Larry: One of the, the, I think where this is really interesting is, in kind of the use case of lead qualification and lead routing.
Larry: Okay, so like there’s this whole bunch of use cases, where it’s like, say you’re selling scooters, okay? Well, you kind of need to determine which is the right scooter for that individual. That you could ask like questions, like you know, what’s your weight, your height, and your gender, you know? And then,
Larry: Like boom, here is the scooter for you. So that’s a difficult experience to represent on an ad but it makes for a more immersive interactive chat experience.
Larry: You know, you can think of the same thing for like you know, software. Sell software, like chat bot software to marketers.
But it’s a very different sell, if we’re actually going to an agency, or an advertiser. Usually an agency is like a different pricing, a different packaging. A different messaging, versus a marketer, who has a different persona.
Larry: And one of the first questions that I ask in my lead ads, or I’m sorry, in my Messenger ads, that I use to sell my product, is I ask them like, you know, “Hey, Jonathan, are you booking at an agency or an in-house marketer?”
Larry: And just by, that’s a very discriminating criteria and depending on how you answer that question, it tells me a lot about how I should redirect the conversation so that it’s pertinent to the customer.
Larry: So it’s just opening up a lot of use cases in advertising that I think previously we were just we were just getting per interests based on kind of the audience criteria.
Larry: Like, oh they have a job title. Or, they’re–
Johnathan: It’s a blanket approach. There’s nothing really that’s unique to the person. Yeah.
Larry: And swapping that out with kind of a user provided conversational criteria, that they’re actually shopping for.
Larry: And these questions can go on forever. Like, I have one customer whose chatbot is like, it’s like, it just stems from GIFs, or sorry, it’s memes of the day, sorry. And people will just go on this chatbot and they’ll request like 46 memes in a row, or something like that.
Larry: You know .
Johnathan: There’s a Guinness Book of World Records just about to happen, in regards to that it sounds like .
Larry: Yeah, so just, it’s a different form of advertising.
Larry: It feels a bit like, on one hand… It’s almost like old fashioned salesmanship. So like, how you used to talk to a salesperson or a sales clerk, like in a mall, or something like that. And the sales representative will try to size you up and try to figure out like how to direct the conversation and how to, you know, build rapport with the person.
This wasn’t really possible with the previous EDM . But now, I think it opens up some of these possibilities.
Johnathan: Totally. And I think a lot of people, one, they don’t really think about that. We work with a company, for example, called BounceX out of New York. And they do, what they basically call, like they create digital sales people, for websites, to guide people to figure out what they want.
And it’s like, you walk into a clothing store and they always ask you, “So, what are you looking for? How can I help you?” And then most of the time you say, “You know, I’m just looking around, I don’t really need any help. Thanks though.” And you browse and you click around on websites.
But, when it’s targeted, and it’s specific and you’re testing the initial questions to get that interaction, it becomes very powerful. Now the trick is then the execution of it, right? Because a lot of people are like, “Oh yeah, I love chatbots. I use them on my website. I use them potentially, you know, if they’re strong enough and smart enough from a marketing perspective, I’ll also use it for that.”
But there’s a lot of ways you can make it and get it wrong. So you know, what would you recommend in that world, from like a recipe perspective, for being able to use this, without getting like, too complex?
Larry: Yeah, look, I don’t think chatbots are complex, right? They’re new.
Larry: So it’s sending out a channel and just like how when you’re starting with email marketing, or Facebook ads, you just kind of have to learn the lingo and you have to kind of understand the objective and stuff like that.
Larry: So, I guess for getting started with chatbots, like you need a chatbot builder. This is not available in Facebook Business Manger, it’s like, they just have an API, it hasn’t been integrated into the ad manager or anything like that.
Larry: It’s, so like MobileMonkey, that’s my chatbot builder, we provide it, it’s a free product.
Larry: You can use that. You know, I think, we talked a little bit about one use case which is, using chatbots, like to create Messenger ads. There’s a couple other common use cases. Another one that’s pretty common, is Chat Blasting, or Messenger Broadcast.
Larry: And so, that’s when you kind of, it’s like a re-engagement strategy. So like all those people who click on your Messenger ads, you get message information basically, for those people.
So they’re added to your base list of Messenger contacts. Like your subscriber list. And so, like one really common thing to do is to set up a drip campaign, so that like in the next say 24 hours or something, you could send them like three, four, five follow up messages.
Like, you know, say 20 minutes later, or an hour later, or two hours later, just to kind of pull them through the user journey a little bit. Like, maybe have them, join your group, or buy a course, or watch the free live webinar.
Larry: Or just like, you make this, this is done on email already.
Larry: These campaigns. But just chat is–
Johnathan: Well the engagement rate, and the open rate on the chatbot side from a Messenger perspective, is quite dramatically higher, most of the time, than email, right?
Larry: Yeah, so email is like five to 10% open rates.
Larry: Pretty typical, 1% click rates. Chat is closer to 80% read rates and 20% click rates.
Johnathan: Wow! I’m gonna give a nugget for that. Keep going, sorry.
Larry: Yeah, this thing, so we’re killing ourselves as marketers that collect these emails and, it’s very difficult to then reconnect with them at scale.
Larry: You know, you don’t like, even if you have their email, it’s like a five or 10% chance of them seeing the next email, you know?
Larry: So, the Messenger, it’s, you get messaging permission from these people, who engage with your bot and you have the ability to send drip campaigns, or even just do periodic chat blasts.
That would be like taking your email list and then saying like, “Hey, what’s up?” And just sending them a broadcast message. That’s another interesting use case.
So, if you wanted to kind of like there’s two kinds of use cases with chatbots. The ad use case, was where the customer was sort of initiating the conversation. They are the one who’s clicking on your ad and then they’re saying like, like they initiated it by clicking on your ad and you’re kind of putting them through this funnel.
The other way is like, you initiate the conversation. So that’s more like email, that’s like where you just blast out a push notification to all these subscribers. You see what I’m saying?
So like, it’s a little more unique, from a marketing perspective, because typically, like, the people who do the email operations are not the same people who do the link in operations.
Larry: You know what I mean?
Larry: But this was a one of these cross functional channels that are both kind of inbound and push modalities, you know?
Johnathan: For sure.
Larry: So, your bot kind of uses these different use cases. Another one is customer support, okay? So that’s like, where you are just trying to answer frequently asked questions. So, using Mobile Monkey, you could install one of those little boxes on the corner of your screen, where you can type in like messages to the company.
And you know using MobileMonkey, like the great thing about this is that every time someone messages your business, they become a subscriber, right? So you get all their contact information and messaging permissions.
You can respond to their questions at scale, like without requiring any human operator.
Larry: It’s kind of like Adworks, where you just specify different responses and keywords for the different responses.
Larry: And then the chatbot kind of analyzes what the user is asking and tries to match, kind of like broad match it, tries to match, the user’s query.
Larry: With the keywords and then responses .
Larry: So if they ask like, “What are your hours?” You know, you’ll have some content for that. But If it doesn’t match, then it doesn’t–
Johnathan: And they can figure that out, like if people have typos and say it in a different, or ask in a different way, sounds like, it’s smart enough to know that?
Larry: Yeah, so. So chatbot builders like MobileMonkey, implement natural language processing out, but it’s basically like broad match. So, synonyms, plurals, you know? Different orders of sort of similar intent.
Like when we analyze what it believes is the intent of the question, and then sort of the most relevant response. It’s kind of like a search engine, or a, like Adworks, basically.
Johnathan: Got it, got it.
Larry: You know?
Johnathan: Okay, so let’s say that you are let’s say they were a software company, for example. Or, you’re doing lead gen and I’m sure there’s always use case for e-commerce. We can talk about that too. Are there different types of chatbots that you would create for different things?
And how, like, how would you go about setting them up? Is there like a thought you have around that? Like, you need one for blasting. You need one for you know, lead generation. Things like that. What are your thoughts?
Larry: So MobileMoney has maybe dozens of different templates based on different types of usages and industries. And for verticals. So like, a use case could be like a survey.
So like one would thing that people do is, they survey the users through messaging rather than email. It’s actually really hard to get people to fill out surveys. Like, IBM, I just did a survey of 5,000 emails and I only got 17 responses. Okay? And so,
Johnathan: I was one of the people
Larry: You know, it’s–
Johnathan: who did not respond. I actually don’t, it might be, the truth actually. I don’t know. Sorry, keep going.
Larry: And then it’s always that one. That you’re like, you’re not my target market. No, just kidding. It’s just, the surveys are another use case for chat, because, you know, just imagine you’re on your mobile phone and you know, you get an email.
And then it said like, “Oh do you wanna fill out this survey?” Like, “No, no freaking way.” Like, you’re not gonna click on this link and wait for this, you know, webpage to load.
Johnathan: It’s like when I call customer service anywhere, they’re like, “You’ve been selected for a five minute survey after this call.” And I’m like, “F that, I’m hanging up as soon as I get whatever I need done.”
Larry: All right. So with messaging, like, remember, you’re like, 100 times more likely to see it in the first place. You know? Like, because emails most people don’t even see.
And then they don’t have to like, you know, open up to a new browser, like all the questions are just buttons to press within messaging, okay? And then they’re prefilled and everything. So, that would be an example of a template that we have in MobileMonkey, where you just fill out the questions.
Larry: And that works. But like, there’s other vertical-wise, kind of chatbot applications, where it’s like, maybe you’re a realtor and you know, like, it’s gonna be the same for every realtor.
They’re gonna, it’s gonna be like, view my listings, it’s gonna be book an appointment. with me to view the listing, or get notified.
Larry: You know, if a new company, I’m sorry, if a new house gets listed in my neighborhood.
Larry: It’s just gonna be the same kind of questions. And same kind of functionality. So we just kind of pre-packaged that and you can kind of just include a link to your chatbot.
Larry: You know, from your business card.
Larry: Or from your email addresses and that can kind of be your kind of like virtual assistant, that’s kind of fields the questions and kind of pre-qualified .
Johnathan: Yeah, I’m thinking of so many things. So this is interesting because the more that you’re speaking about this, the more I think of like, one of the recipes that we use quite often on our landing pages.
We call it the bread crumb technique and it basically is, initially we ask questions that are anonymous for the visitors, like we don’t you know, wanna know their name. We don’t want their phone number. We don’t want their email. And they answer these questions that are related to their problem, that we wanna help solve.
And then on the second page, we basically try to book an appointment if it’s lead gen for a client, or even for ourselves. So when you’re saying this, what I’m like, the things popping up in my head, like the realtor example.
You can start off by asking like, “Are you looking to buy or sell?” And if either, or, you can even say, “Well how many bedrooms are you looking for? How many bathrooms? Is it one story or is it two story?”
And you can get them to kind of commit to this detail where at the end, they’re actually more likely to wanna talk to somebody if you are doing lead gen, versus just hitting them straight away with like that ad for example, that you, that we’re comparing it to, where it just says, leave your name, email and phone number.
There’s not a lot of value or excitement, or you’re not getting their foot in your door as well, it seems like.
Larry: That’s why those chat pop up windows work so well.
Larry: People view it as sort of a lower friction like, as compared to filling out a form.
Larry: Like filling in a form is like, “Oh my God. I’m gonna be added to this list.” and,
Johnathan: I’m gonna have to give a blood sample. And I’m gonna be screwed, basically, because 10 sales people are gonna call me. And that’s what’s gonna happen.
Larry: All right. So yeah, but the funny thing about the chatbots is that we already have their information.
Johnathan: Right. They don’t know that .
Larry: Like, as soon as they say hello.
Larry: Okay? As soon as they say hello or anything. Well, the business is then able to collect the user’s information, like, first name, last name. You know, location, time zone, gender, profile photo.
Johnathan: Blood type.
Larry: Et cetera.
Johnathan: just kidding.
Larry: Yeah. That information is provided automatically.
Larry: In order to customize the conversation, so that like I say, “Hey Jonathan, did you think about this bungalow?” Or something like that, you know, I can customize the conversation to reflect that
Larry: Kind of their identity. You know?
Johnathan: Yep. No, it makes a ton of sense. And so, I mean same thing for e-commerce, if you’re selling products. It can guide the person to the correct product. It’s just like that digital sales person.
As people define their audiences, define either using a template, or building, I guess, their own logic, are there some specific like, you know, like don’t do these things, like it’s common and this a mistake that happens frequently. Are there any of those?
Larry: The thing that you need to be worried about, is your block rate.
Larry: Okay? So, block rates are kind of the equivalent of like an email un-subscription rate.
Larry: I don’t know if you know this, but if you, if you use like Mailchimp or something and if you have more than a 1% complaint rate. Like people saying, “I never signed up for this thing in the first place.” You’ll actually be banned from the platform.
Larry: Okay? Like you’ll have to, you’ll have to take your email list and go to some other provider.
Johnathan: Go to a shadier provider. Just kidding.
Larry: With lower standards. Well, you know, it’s sad, but true.
Larry: The challenge here with messaging is, if you have like a six, seven, 8% block rate, so if people actively saying like, “Cut it out. Stop messaging me ever.” Okay? Then you can actually lose messaging permissions. So like, the ability to do all these automated responses and questions .
Larry: Deactivated from your page and so you’ve got to be monitoring your block rate.
You know, so, you were just asking like, are there any things, mistakes that people make? The mistake they make is, they don’t, they’re not monitoring the unsubscribes, no, no, not the unsubscribes. Unsubscribes are fine.
Larry: Okay, that doesn’t count, because they’re not blocking you.
Larry: But it’s the block rate. The strategy is not to try to keep these people on your list forever and have no way to get out. It’s actually the opposite, you wanna, curate kind of a list of super fans.
Larry: Something like where the people who were on your list, really, really wanna be there because the content is so great, or because, you know, it’s so useful. And so, this implies like, like when I send out a message blast, I put the end subscription messages at the beginning of the message, not at the end of the message.
Johnathan: Really. That’s crazy!
Larry: Well, I have one and there can be no ambiguity as to how to unsubscribe.
Johnathan: Right. Because unsubscribe doesn’t matter. Because you’re talking like, the block rate, right? So, you’d rather have them feel like they’re able to do that first, and then not have it turn into an actual block, is that correct?
Larry: So like my block rate is less than 1%.
Larry: And it’s because we do a couple of things. We do scheduling. So like, you know, we’re not gonna send out a blast at like two a.m. or something, you know? We have your time zone because, you know, Facebook gives you the time zone of the contacts. And we schedule them to be at a specific time relative to their time zone.
Larry: Like 8 a.m. or 10 a.m. or whatever. So and then making it very easy to unsubscribe. So like, some of the big mistakes that these e-commerce companies were doing, were they were just, like e-commerce is among the most scammy like in terms of email marketers, if you look through your email inbox.
Larry: Like,like that’s like a vertical that is pretty awful with sending out like you know, weekly deals, or daily deals, or daily reminders.
Like it’s just all, a buy now kind of stuff and like I think, I think messaging is a different medium. Like you can’t, it’s not just like a, it’s not just copy and paste sending out in email.
Larry: We really need to keep the block rates low and engagement rates higher, otherwise, we’ll get booted off this platform.
Larry: And it’s not like you can pack your bags, you know, and go to like, I don’t know.
Johnathan: There’s no other platform! Like, you’re screwed!
Larry: Yeah. Exactly. So that’s a big mistake.
Larry: You gotta think about the value that we’re providing and it’s not for you to decide that this is valuable. It’s based on your block rate, you know?
Johnathan: Makes sense.
Johnathan: Makes sense.
Larry: It’s kind of like a crowdsource, the determination of your value.
Johnathan: Like who could ask for a better feedback loop? Like you better take you know, you better listen to what they have to say and see it and then make an adjustment because these are your prospects, in a sense. So, it seems like a no-brainer.
Larry: I think, I think this is so great. Because like if the, marketers get a little nervous when I describe this, but actually, it’s really wonderful, because it means that what happened to email, like how it’s like, you know, five or 10% open rates,
Larry: And all this stuff. Like, it’s, I’m not gonna say never gonna happen to chat, but, it’s gonna take a while, it’s gonna be a much longer kind of cycle, you know? Because the bad actors kind of automatically get eliminated.
Johnathan: Booted pretty quickly.
Johnathan: Yeah. I mean, and like the saying goes, “Marketers ruin everything.” And so, it could be a matter of time, but we’ll see. So far, it seems like a great opportunity to take advantage of.
Is there anything, as we wrap up the conversation, Larry, that like, for me, everything that I’m hearing, and I’m not just saying this because I’ve like, basically, I’m doing product placement for you, because I know the value of MobileMonkey.
Is there anything that we wanna tell the listeners, in regards to like what they should be aware of? Because like, we talked about the audience. We talked about the templates. We talked about things that you can do yourself, in regards to getting direct response value and out performing a lot of other ad types. Anything we’re missing?
Larry: Yeah. It’s just the future of messaging, I think. And so, it’s really an exciting and emerging field here.
If you go to China today or the other eastern nations, like, there’s no newsfeed, okay? People, like, if we chat, so like, or other chat application to order their taxis or, order their lunch, or,
Johnathan: I heard people do like, pay their mortgages through that too. I haven’t seen it, but I’ve heard about it.
Larry: Yeah so, the is a very strange construct. It’s a very round about way of like, like this, your messaging contact list makes more sense, like it’s like okay, I wanna message a business, you know?
Like, I’m gonna just kind of scroll through this random you know thing of newsfeed and maybe I’ll, you know, see a business ad that I’m gonna respond to.
Larry: It’s a very unusual construct.
Larry: And newsfeed. And Facebook knows that it’s not the future of Facebook.
Larry: Like they, it’ll sort of fade and it’s like every single session was about messaging and,
Larry: You know, the future of Facebook is kind of this smaller, and private conversations.
Johnathan: People on our team, were at the Google Marketing Live event and one of the things that stood out, that I was really curious about was like, they’re having conversational display ads.
That they’re gonna start rolling out, which is kind of a very similar to like what we’re talking about too. Like they’re not, Facebook is not the only company knowing that there’s a lot of advantage to this. So that’s really encouraging to hear.
Larry: Yeah, so and there was a link in the New York Times saying that Facebook is planning on integrating Whatsapp and Instagram and Messenger. It’s like the number one kind of objection to like, you know, doing marketing on Messenger is like, not everyone uses Messenger.
Larry: And it’s like, you know, only a couple billion people. But like, who doesn’t have a Whatsapp or an Instagram, or a Messenger? And if they don’t have any of those, is that really a valuable lead that we’re talking to, you know?
Johnathan: True, true.
Larry: So like, it’s just, it’s one of these really neat kind of inflection points, in the technology adoption curve.
Larry: And you know, I expect to see like, it’s very un-used right now. It’s like, you know, probably 1% of businesses use messaging. But, I think we would not be crazy to see like 30, 40% of them, using this stuff in the next three to five years.
Johnathan: That’s why we’re talking. Exactly.
Larry: Everything I know about marketing is like you have to jump on channels quickly.
Larry: Like you know, it’s like, of my million Twitter followers, like, 80% of them were in the first two years. There was nobody on the platform.
Larry: You know? It’s just, it’ll never be to use these systems and we really get a head start, in terms of like, you know.
Larry: And et cetera.
Johnathan: So you’re basically saying that I should not prioritize my radio advertising plans, over the Facebook Messenger plans that I also have?
Larry: Yeah, it just requires a little bit of foresight in where the industry is going.
Johnathan: For sure.
Larry: That’s the one thing.
Johnathan: Very kind way of you to explain that, in a nice way, to shut me down, I appreciate that.
Well Larry, this has been super, super fun. I know a ton of people are gonna get a value, because not only is the conversation between you and I gonna be in video format, audio format, and also written, we’re also gonna share the webinar that we did, which gives a whole nother, dimension to this of what you can do, beyond this.
So, I wanna thank you for your time. And you know, let you enjoy the rest of your day.
Larry: Awesome, thanks Jonathan. Thanks everyone for joining.
Johnathan: Of course. Awesome, man, thank you.
Chatbot Mastery With Johnathan Dane
Johnathan: So, this was a very interesting episode because I have done a lot of reading, a lot of research when it comes to chatbots in general, obviously you guys had the pleasure of listening to our guests about this too.
One thing I wanted to help everybody out with, that we’ve seen internally from our own experimentation, but in addition to that, our clients that we’re working with too, is that, just like any form, any flow on a website, chatbots get hype because of, like, their engagement rate.
And the thing that I think frequently about is the same thing that, like, SMS and text is being touted in the marketing world for having a high open rate. It was the same thing with email, when email was newer too, not that SMS or text is newer, but eventually, us marketers and us advertisers are gonna ruin that channel as well too.
So when we’ve had the opportunity to talk to so many guests on this show so far, with so many different interpretations and opinions, one thing you’ve got to keep in mind is that there’s another human on the other end of that screen. Whether they’re interacting with your chatbot on desktop, mobile, tablet, it actually doesn’t really matter.
One thing you have to make very, very certain of to make sure it’s successful, not just for the chatbot to do an activity, but for you ultimately to make more money, which I think is a lot of people’s goals in the back of their mind when it comes to marketing or advertising in general.
It’s not to get open rates, it’s not to get reply rates, it’s not to get booking rates or demo rates, or MQLs to SQLs, those are all indicators, those are all leading indicators for what’s more important, which is actually the sale.
So when you do things and you’re testing things and you’re listening to a podcast like this, or reading articles about other things, one thing you need to keep in mind is, first off, you may never have enough data to actually know what you’re testing. Meaning, does doing this chatbot open message versus another one actually have a better close rate that turns into making more money, you need to have a lot of volume before you can figure that out.
And, when people actually do the demo call, let’s say you’re using Drift for example, and you are trying to schedule demos, it is as much the intent of the person who is on the demo, who’s your prospect, and the deliverability and the attitude and the tone and the energy and the problem solving of the product on that demo that has an impact, so you’re never really fully running a perfect experiment.
So, the best advice that I can leave you off with is make sure that your chatbot gives value. Meaning, it gives people and points them in the right direction, it doesn’t do any bait and switch.
If you do have to ask a person to grab a time on a calendar to book a meeting, make sure you explain why you can give the answers that they’re looking for right then and there.
Let’s say for example that you are a lawyer, and you’re, in fact, a personal injury lawyer. We’ve used this example before too. Your demobot, or like the free consultation chatbot that you’re using, can give a person more intel in regards to the questions that the lawyer would ask the prospect, if they’re on the phone call with them.
Like, well, when did the accident happen? What kind of accident happened? Are you in pain right now? Did you go see a doctor yet? Are you taking any medicine for it yet?
Those kind of things help the prospect know that they’re getting closer and closer to a solution, versus just saying, you know, choose your path through the chatbot and then eventually book a time.
Or another example could be like, grab an ebook. They might want that grab that ebook and they might not, so, like again, think of this as a human being. Like you’re a receptionist in your own office, and people are coming in, which is the equivalent of your website or your laying page, and they’re engaging with you, the chatbot.
Here in this situation, you’re the receptionist, and you’re asking them questions that point them in the right direction. You’re not just giving them one stop shop and then be happy with it. Because a lot of times, the hope is to automate and simplify, but we as humans, just, that only really works if we can get answers to our own questions quicker.
So, keep that in the back of your mind as you test that out.